Dear All,
I'm new to Stata and not an expert in statistics, currently doing a research using path analysis with SEM Stata. I'm using secondary data and there are 2 exogenous variables, 1 intervening, 1 endogenous and 5 control variables. There are 3 things that's confusing me:
1. Do I need to test classical assumptions (normality, multicollinearity, heteroskedasticity, and autocorrelation) if using SEM?
2. If I have 5 control variables, do I need to put those in the path diagram? Cos if I put those control variables in path diagram, my model would become too complex.
3. Is goodness of fit (GoF) a mandatory test in SEM?
Sorry if my question is too basic as I've never using Stata and this is my second research to complete my master program. Thank you.
I'm new to Stata and not an expert in statistics, currently doing a research using path analysis with SEM Stata. I'm using secondary data and there are 2 exogenous variables, 1 intervening, 1 endogenous and 5 control variables. There are 3 things that's confusing me:
1. Do I need to test classical assumptions (normality, multicollinearity, heteroskedasticity, and autocorrelation) if using SEM?
2. If I have 5 control variables, do I need to put those in the path diagram? Cos if I put those control variables in path diagram, my model would become too complex.
3. Is goodness of fit (GoF) a mandatory test in SEM?
Sorry if my question is too basic as I've never using Stata and this is my second research to complete my master program. Thank you.
